This essay was submitted to Open Philanthropy's Cause Exploration Prizes contest (before the deadline). We are uploading some entries late, but all good-faith entries were considered for prizes.
Open Philanthropy should officially start up a meta-science cause area.
It has toyed with meta-science for many years. Indeed, back in 2012-13, Holden Karnofsky gave me a one-pager idea he had seen from Steve Goodman and John Ioannidis at Stanford (which I helped them develop into METRICS, the Meta-Research Innovation Center at Stanford).
But it’s time to make this a full-fledged cause area.
The case is simple:
Scientific innovation is a key driver of economic and human flourishing.
But our current scientific ecosystem is too tied down by bureaucracy, incrementalism, status quo biases, poor reproducibility, and more. In a world so highly affected by scientific outliers, we need to be obsessed with reducing upside decay.
We may not be able to predict how to optimize for outliers, but we can at least stop doing all of the things that prevent outliers. After all, if we drive even one Katalin Karikó out of the system, that could cost us a trillion dollars.
Let’s take just one indicator: whether we’re getting the maximum value from talented scientists who happen to be younger.
Einstein had his best year at age 26. Heisenberg published his uncertainty principle at age 26. Paul Samuelson’s masterpiece “Foundations of Economic Analysis” was based on work he completed at age 26. James Watson was 25 when he helped discover the structure of DNA.
Yet in today’s scientific workforce, a 26-year-old might be merely half-way through a doctorate, and would virtually never be eligible for a major scientific grant. Indeed, the median age for a first NIH grant is 42 for men and 44 for women, meaning half of such grants are given to people above that age!
Such a lengthy timeline, combined with low pay in early years and the uncertainty of getting tenure, almost certainly discourages many highly-talented people from entering science in the first place.
(Imagine telling Einstein or Heisenberg at age 25 that they would have to wait 10 more years before they were deemed worthy of scientific recognition, but that they could earn many times more money in tech or finance.)
On top of that, there is reason to think that even when smart and talented people get doctorates in science, the best among them can be selected out of the academic workforce, because they want to spend time on the most important problems at the forefront of a field rather than cranking out marginal publications. The hard problems, after all, might take 8 or 10 years to make progress, and no one has time for that.
If we could fix this problem alone, the future improvements to scientific and economic progress might be in the trillions. I can’t in good conscience create a lengthy spreadsheet calculation, because nearly every number would be made-up. The more important point in any event is whether one agrees with the overall intuition. (If you don’t agree, then no spreadsheet would convince you, while if you do agree that radical scientific progress is possible and that current practices discourage it, no spreadsheet is necessary.)
So far, I’ve just been addressing the problem of how the system wastes young talent or drives it away altogether. There are many other problems to address, such as the lack of reproducibility, the inherent biases and conservatism of peer review, the lack of experimentation with different models of funding, the micromanagement and overly-bureaucratic nature of the funding system, and more.
Suffice it to say: Our scientific system has found many ways to stifle curiosity and innovation. A meta-science effort to promote better policies, while carefully studying what works and what doesn’t, could ultimately create benefits so staggering that they are equivalent to developing the steam engine or discovering the structure of DNA.